Exploring Global Happiness: A Comprehensive Analysis
of the Impact of Socio-Economic Factors and Health on Happiness Scores
Across Regions
Prakruthi Harish
2024-08-19
Section 1: Background
This project addresses the crucial question of what drives global
happiness, moving beyond traditional economic measures like GDP. It
focuses on how various socio-economic factors influence the Life Ladder
score, a key indicator of happiness, and how these influences differ
across regions and countries. Understanding these dynamics is vital for
policymakers aiming to improve citizens’ quality of life. Using data
from the World Happiness Report, the project analyzes variables such as
GDP per capita, social support, life expectancy, freedom of choice,
generosity, and perceptions of corruption.
By exploring these factors, the project highlights patterns and
correlations that explain variations in happiness globally, offering
data-driven insights for policy interventions. Helliwell et.al (2019) in
the World Happiness Report emphasized the impact of economic and social
factors on happiness. This project expands on that analysis by
incorporating a broader range of economic measures and offering a
multi-regional perspective. By examining these factors over time, the
project provides deeper insights into the trends that shape global
happiness.
The primary objective is to understand the correlation between
various factors and their impact on happiness, identifying
region-specific trends and global patterns. The project employs
exploratory data analysis (EDA), using statistical tests and
visualizations like histograms, Q-Q plots, box plots, and interactive
dashboards. These findings offer actionable solutions for enhancing
well-being worldwide.
Section 2: Methods
This project analyzes two key datasets from the World Happiness
Report: one covering 2005-2023 for a broad temporal view, and another
focused on 2024 for current socio-economic insights. These datasets,
sourced from the World Happiness Report and Kaggle, are vital for
understanding both short-term and long-term global well-being. The data,
gathered through the Gallup World Poll, includes metrics like GDP per
capita, social support, life expectancy, and more, across 150+
countries. The 2024 dataset contains 143 rows and 12 columns, while the
2005-2023 dataset has 2363 rows and 11 columns.
2.2. Data Cleaning and Pre-processing
It is essential to ensure that the data is clean and ready for
exploration before starting the analysis. So the pre-processing step
involved addressing missing values, non-ASCII characters, and other
potential data inconsistencies. Missing values were handled by imputing
them with the mean of the respective columns. This approach ensured that
the data remained representative and accurate. Additionally, non-ASCII
characters were removed from the country names to avoid errors during
analysis and visualization. These pre-processing steps were crucial to
maintain the integrity of the dataset and to ensure reliable
results.
2.3. Exploratory Data Analysis (EDA)
The first step in the analysis carried out in this project was to
explore the distribution of key variables, including Ladder score, GDP
per capita, and Social Support. This was accomplished by visualizing the
data through histograms, which provided initial insights into the
overall trends and highlighted any anomalies or patterns in the data.
However, while histograms provided a general sense of the data
distribution, they did not offer detailed insights into the normality of
the data. To address this, Q-Q plots were generated to more assess
whether the variables followed a normal distribution. This step was
crucial for determining the suitability of parametric statistical tests
in subsequent analyses. However, Q-Q plots alone did not reveal how
these distributions varied across different regions. In order to fill
this gap box plots were employed to examine the regional distribution of
Ladder scores within the 2024 dataset, offering a more granular view of
how happiness varied across different regions. To further contextualize
these findings, a pie chart was created to visualize the distribution of
countries by region. This step clarified the sample size for each
region, thereby enhancing the interpretation of the box plot findings.
While this added a layer of understanding about regional representation,
it still didn’t explain how factors like health might influence
happiness.
Further analysis focused on the relationship between health and
happiness, with the average Healthy Life Expectancy across regions being
plotted to explore how disparities in health might influence regional
happiness levels. An interactive choropleth map was then created to
track changes in happiness scores over time. This dynamic approach
allowed for a comprehensive understanding of global happiness trends
over the years. However, this broader view needed to be complemented by
a more focused analysis. Thus, a comparative analysis was conducted to
identify the happiest and unhappiest countries based on their cumulative
Ladder scores. This helped pinpoint specific countries that consistently
ranked at the top or bottom of the happiness spectrum. This step was
followed by statistical validation through normality checks and T-tests.
The correlations between various socio-economic factors and happiness
were analyzed, both at the country and regional levels. To visualize the
relationships between the Ladder score and various socio-economic
factors, scatter plots with regression lines were created which
confirmed the linearity and strength of these relationships. This
provided a detailed view, yet to enhance the depth of analysis, an
advanced interactive visualization was created. This allowed for a more
granular exploration of how different factors influence happiness across
regions. To further deepen the analysis, an advanced interactive
visualization was created. This allowed for a more granular exploration
of how different factors influence happiness across regions, offering a
more nuanced understanding of the complex dynamics at play.
The code to implement each of these methods, along with the results
obtained, is presented below. The results include a series of
visualizations and analyses that provide insights into the key drivers
of happiness across the globe.
Note: Please scroll down the Slidey page to explore the results and
detailed analysis.
Section 3: Implementation
and Results
3.1. Install and load libraries
This step loads all the packages required for exploratory data
visualization and analysis conducted in this project.
##
## The downloaded binary packages are in
## /var/folders/9c/v_4gdjk147ndq15wh3wpcpy80000gp/T//RtmpeJiYG6/downloaded_packages
##
## The downloaded binary packages are in
## /var/folders/9c/v_4gdjk147ndq15wh3wpcpy80000gp/T//RtmpeJiYG6/downloaded_packages
##
## The downloaded binary packages are in
## /var/folders/9c/v_4gdjk147ndq15wh3wpcpy80000gp/T//RtmpeJiYG6/downloaded_packages
##
## The downloaded binary packages are in
## /var/folders/9c/v_4gdjk147ndq15wh3wpcpy80000gp/T//RtmpeJiYG6/downloaded_packages
##
## The downloaded binary packages are in
## /var/folders/9c/v_4gdjk147ndq15wh3wpcpy80000gp/T//RtmpeJiYG6/downloaded_packages
##
## The downloaded binary packages are in
## /var/folders/9c/v_4gdjk147ndq15wh3wpcpy80000gp/T//RtmpeJiYG6/downloaded_packages
##
## The downloaded binary packages are in
## /var/folders/9c/v_4gdjk147ndq15wh3wpcpy80000gp/T//RtmpeJiYG6/downloaded_packages
##
## The downloaded binary packages are in
## /var/folders/9c/v_4gdjk147ndq15wh3wpcpy80000gp/T//RtmpeJiYG6/downloaded_packages
## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr 1.1.4 ✔ readr 2.1.5
## ✔ forcats 1.0.0 ✔ stringr 1.5.1
## ✔ ggplot2 3.5.1 ✔ tibble 3.2.1
## ✔ lubridate 1.9.3 ✔ tidyr 1.3.1
## ✔ purrr 1.0.2
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag() masks stats::lag()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors
##
## Attaching package: 'kableExtra'
##
##
## The following object is masked from 'package:dplyr':
##
## group_rows
##
##
## Loading required package: SnowballC
##
## corrplot 0.94 loaded
##
##
## Attaching package: 'plotly'
##
##
## The following object is masked from 'package:ggplot2':
##
## last_plot
##
##
## The following object is masked from 'package:stats':
##
## filter
##
##
## The following object is masked from 'package:graphics':
##
## layout
##
##
##
## Attaching package: 'reshape2'
##
##
## The following object is masked from 'package:tidyr':
##
## smiths
3.2. Load the Datasets
This step imports the datasets and provides a preliminary
understanding of structure and content of data. This step helps identify
data types and certain variable format inconsistencies in the datasets
early in the process.
## The 2024 dataset has 143 rows and 12 columns.
## The 2005-2023 dataset has 2363 rows and 11 columns.
## 'data.frame': 143 obs. of 12 variables:
## $ Country.name : chr "Finland" "Denmark" "Iceland" "Sweden" ...
## $ Regional.indicator : chr "Western Europe" "Western Europe" "Western Europe" "Western Europe" ...
## $ Ladder.score : num 7.74 7.58 7.53 7.34 7.34 ...
## $ upperwhisker : num 7.82 7.67 7.62 7.42 7.41 ...
## $ lowerwhisker : num 7.67 7.5 7.43 7.27 7.28 ...
## $ Log.GDP.per.capita : num 1.84 1.91 1.88 1.88 1.8 ...
## $ Social.support : num 1.57 1.52 1.62 1.5 1.51 ...
## $ Healthy.life.expectancy : num 0.695 0.699 0.718 0.724 0.74 0.706 0.704 0.708 0.747 0.692 ...
## $ Freedom.to.make.life.choices: num 0.859 0.823 0.819 0.838 0.641 0.725 0.835 0.801 0.759 0.756 ...
## $ Generosity : num 0.142 0.204 0.258 0.221 0.153 0.247 0.224 0.146 0.173 0.225 ...
## $ Perceptions.of.corruption : num 0.546 0.548 0.182 0.524 0.193 0.372 0.484 0.432 0.498 0.323 ...
## $ Dystopia...residual : num 2.08 1.88 2.05 1.66 2.3 ...
## 'data.frame': 2363 obs. of 11 variables:
## $ Country.name : chr "Afghanistan" "Afghanistan" "Afghanistan" "Afghanistan" ...
## $ year : int 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 ...
## $ Life.Ladder : num 3.72 4.4 4.76 3.83 3.78 ...
## $ Log.GDP.per.capita : num 7.35 7.51 7.61 7.58 7.66 ...
## $ Social.support : num 0.451 0.552 0.539 0.521 0.521 0.484 0.526 0.529 0.559 0.491 ...
## $ Healthy.life.expectancy.at.birth: num 50.5 50.8 51.1 51.4 51.7 ...
## $ Freedom.to.make.life.choices : num 0.718 0.679 0.6 0.496 0.531 0.578 0.509 0.389 0.523 0.427 ...
## $ Generosity : num 0.164 0.187 0.118 0.16 0.234 0.059 0.102 0.078 0.04 -0.123 ...
## $ Perceptions.of.corruption : num 0.882 0.85 0.707 0.731 0.776 0.823 0.871 0.881 0.793 0.954 ...
## $ Positive.affect : num 0.414 0.481 0.517 0.48 0.614 0.547 0.492 0.491 0.501 0.435 ...
## $ Negative.affect : num 0.258 0.237 0.275 0.267 0.268 0.273 0.375 0.339 0.348 0.371 ...
## Country.name Regional.indicator Ladder.score upperwhisker
## 1 Finland Western Europe 7.741 7.815
## 2 Denmark Western Europe 7.583 7.665
## 3 Iceland Western Europe 7.525 7.618
## 4 Sweden Western Europe 7.344 7.422
## 5 Israel Middle East and North Africa 7.341 7.405
## 6 Netherlands Western Europe 7.319 7.383
## lowerwhisker Log.GDP.per.capita Social.support Healthy.life.expectancy
## 1 7.667 1.844 1.572 0.695
## 2 7.500 1.908 1.520 0.699
## 3 7.433 1.881 1.617 0.718
## 4 7.267 1.878 1.501 0.724
## 5 7.277 1.803 1.513 0.740
## 6 7.256 1.901 1.462 0.706
## Freedom.to.make.life.choices Generosity Perceptions.of.corruption
## 1 0.859 0.142 0.546
## 2 0.823 0.204 0.548
## 3 0.819 0.258 0.182
## 4 0.838 0.221 0.524
## 5 0.641 0.153 0.193
## 6 0.725 0.247 0.372
## Dystopia...residual
## 1 2.082
## 2 1.881
## 3 2.050
## 4 1.658
## 5 2.298
## 6 1.906
## Country.name year Life.Ladder Log.GDP.per.capita Social.support
## 1 Afghanistan 2008 3.724 7.350 0.451
## 2 Afghanistan 2009 4.402 7.509 0.552
## 3 Afghanistan 2010 4.758 7.614 0.539
## 4 Afghanistan 2011 3.832 7.581 0.521
## 5 Afghanistan 2012 3.783 7.661 0.521
## 6 Afghanistan 2013 3.572 7.680 0.484
## Healthy.life.expectancy.at.birth Freedom.to.make.life.choices Generosity
## 1 50.5 0.718 0.164
## 2 50.8 0.679 0.187
## 3 51.1 0.600 0.118
## 4 51.4 0.496 0.160
## 5 51.7 0.531 0.234
## 6 52.0 0.578 0.059
## Perceptions.of.corruption Positive.affect Negative.affect
## 1 0.882 0.414 0.258
## 2 0.850 0.481 0.237
## 3 0.707 0.517 0.275
## 4 0.731 0.480 0.267
## 5 0.776 0.614 0.268
## 6 0.823 0.547 0.273
3.3. Visualize data distributions
Visualize the data with histogram to understand the distribution of
key variables like Ladder score, GDP per capita, Social Support, etc.,
in both the datasets. This step provides a summary of the general trends
in the data through histogram distribution plots.
Analysis: The histograms provide a visual summary of the key
variables from the two datasets, reflecting the overall trends and
distributions. From the histograms it can be ascertained that, the
distribution of ladder score is roughly symmetric which implies that,
very few countries have extremely high or extremely low ladder scores in
both the datasets. However, the histograms of Log GDP per Capita of both
the datasets are observed to be right-skewed (positively skewed) which
highlights that most countries have been in the lower GDP per capita
range, with very few countries being significantly wealthy.
In summary, the two datasets reveal trends in global well-being, with
2024 reflecting relatively high scores in happiness, social support, and
life expectancy. However, over the longer period from 2005-2023, there’s
more variability, especially in social support and life expectancy,
suggesting that global trends have fluctuated over time, possibly due to
varying global events and changes. This variability could be attributed
to different global events and changes over the years, highlighting the
importance of understanding these trends in the context of the data.
However, the nature of non-normality seen in certain variables like, Log
GDP per capita or Life expectancy will be further investigated in the
next steps.
3.4. Data cleaning
This step identifies and addresses any missing values in the
datasets. This is because missing values can distort the observed
patterns and trends. Therefore, it’s essential to identify and address
missing data to maintain the integrity of this Exploratory Data Analysis
(EDA).
## Log.GDP.per.capita Social.support Healthy.life.expectancy
## 1 3 3 3
## Freedom.to.make.life.choices Generosity Perceptions.of.corruption
## 1 3 3 3
## Dystopia...residual
## 1 3
## Log.GDP.per.capita Social.support Healthy.life.expectancy.at.birth
## 1 28 13 63
## Freedom.to.make.life.choices Generosity Perceptions.of.corruption
## 1 36 81 125
## Positive.affect Negative.affect
## 1 24 16
3.5. Handle missing values
To handle missing values, ‘NAs’ and ‘NaNs’ are replaced with mean of
the respective columns to ensure data integrity and consistency for
further analysis. Here, ‘grouping’ ensures that calculating the ‘mean’
and filling in missing values are done within the context of each
country, maintaining the integrity of country-specific data. And
‘ungrouping’ allows the ‘mean’ to be applied across the entire dataset,
which is necessary for filling in any gaps left by the grouped
operation.
## # A tibble: 6 × 12
## Country.name Regional.indicator Ladder.score upperwhisker lowerwhisker
## <chr> <chr> <dbl> <dbl> <dbl>
## 1 Finland Western Europe 7.74 7.82 7.67
## 2 Denmark Western Europe 7.58 7.66 7.5
## 3 Iceland Western Europe 7.52 7.62 7.43
## 4 Sweden Western Europe 7.34 7.42 7.27
## 5 Israel Middle East and North Afr… 7.34 7.40 7.28
## 6 Netherlands Western Europe 7.32 7.38 7.26
## # ℹ 7 more variables: Log.GDP.per.capita <dbl>, Social.support <dbl>,
## # Healthy.life.expectancy <dbl>, Freedom.to.make.life.choices <dbl>,
## # Generosity <dbl>, Perceptions.of.corruption <dbl>,
## # Dystopia...residual <dbl>
## Country.name year Life.Ladder Log.GDP.per.capita Social.support
## 1 Afghanistan 2008 3.724 7.350 0.451
## 2 Afghanistan 2009 4.402 7.509 0.552
## 3 Afghanistan 2010 4.758 7.614 0.539
## 4 Afghanistan 2011 3.832 7.581 0.521
## 5 Afghanistan 2012 3.783 7.661 0.521
## 6 Afghanistan 2013 3.572 7.680 0.484
## Healthy.life.expectancy.at.birth Freedom.to.make.life.choices Generosity
## 1 50.5 0.718 0.164
## 2 50.8 0.679 0.187
## 3 51.1 0.600 0.118
## 4 51.4 0.496 0.160
## 5 51.7 0.531 0.234
## 6 52.0 0.578 0.059
## Perceptions.of.corruption Positive.affect Negative.affect
## 1 0.882 0.414 0.258
## 2 0.850 0.481 0.237
## 3 0.707 0.517 0.275
## 4 0.731 0.480 0.267
## 5 0.776 0.614 0.268
## 6 0.823 0.547 0.273
3.6. Normality Check Using Q-Q Plot
The histogram visualization performed in Step 3 gave a quick sense of
how data is spread across different values and showed how certain
varibles deviated from the normal distribution. However, to specifically
understand where the deviations occurs, a Normality check using Q-Q plot
is performed in this step. These Q-Q plots serve as a follow-up to
confirm the normality or non-normality suggested by the histograms. This
step is a validation for the analysis made in Step 3.
Analysis: The Q-Q plots show that Ladder Scores in both the datasets
mostly follows the red line (theoretical quantile) but there are slight
deviations at both the ends (tails). This suggests a mostly normal
distribution. Further, for GDP per capita, the Q-Q plots of both the
datasets show some deviation from the line in the lower tail, which
indicates the presence of lower-than-expected GDP values that don’t
align with the normal distribution. In summary, the Q-Q plots confirmed
the findings from the histograms, providing a detailed visualization of
how some variables in the data closely align with a normal distribution
while few others seem to have a slight deviation from the normal (mostly
right skewed).
3.7. Box Plot Analysis
In the previous step, the Q-Q plots provided an overall assessment of
the distribution of Ladder Score and other variables in the datasets.
However, they did not offer insights into how these distributions differ
across specific regions. Therefore, Box plots are employed in this step
to specifically visualize the distribution of the Ladder Score across
different regions in the 2024 dataset. This transition from a general
examination of distributional normality to a more detailed investigation
allows for the identification of regional variations and any outliers
within the distribution.
Analysis: From the above boxplot visualization, we can identify that
Western Europe has the highest ladder score and Sub-Saharan Africa has
the lowest ladder score. Further, we can also identify that Middle East
and North Africa has the largest spread in the distribution of the
ladder score. On the contrary, North America and ANZ has the lowest
spread. However, it is essential to understand the underlying structure
of the data in terms of how many countries are contributing to each
region’s score.
3.8. Distribution of Countries by Region
The Box plots in the previous step, gave insights into the spread and
central tendency of the Ladder Scores across regions, but they did not
reveal whether these observations were based on a large or small number
of countries, which could potentially affect the reliability and
generalizability of the findings.
By generating a Pie Chart that shows the distribution of countries by
region, it is easier to visualize the number of countries represented in
each region. This step adds context to the Box plot analysis by
revealing the size of the sample from each region and providing an
understanding of whether the observed trends in the Box plots are based
on a large or small number of data points.
Analysis: The pie chart reveals that Sub-Saharan Africa constitutes
the largest proportion of countries among all regions, accounting for
approximately 24.5% of the total sample. When this is correlated with
the box plot findings, which showed that Sub-Saharan Africa has the
lowest Ladder Scores, it suggests that a significant portion of the
global dataset is represented by countries in this region, many of which
report lower happiness scores. Likewise, Western Europe is represented
by a considerable portion of the countries (14%) in the dataset, as
shown in the pie chart. This substantial representation, along with the
high Ladder Scores observed in the box plot for Western Europe in the
previous step, suggests that the positive trends in happiness in this
region are not isolated to a few countries but are instead reflective of
a broader regional trend. However, the pie chart reveals that North
America and ANZ have the lowest representation in the dataset, making up
only 2.8% of the total countries. This region displayed a low spread in
Ladder Scores in the previous step, which indicates a relatively
consistent level of happiness within the region. However, the small
sample size, as shown in the Pie Chart, suggests that these findings may
require to be interpreted with caution.
3.9. Mean Healthy Life Expectancy Across Regions
In the previous steps, the analysis of the distribution of Ladder
Scores and the number of countries per region provided insights into the
overall happiness and its variability across different regions. However,
happiness could be influenced by various factors, including health. So,
to gain a deeper understanding of the regional differences in
well-being, it is essential to examine ‘Healthy Life Expectancy at
birth’, which is a critical component of overall quality of life and
well-being. Healthy life expectancy is calculated as the average number
of years a newborn infant would live in good health. The plot of ‘Mean
Healthy Life Expectancy across regions’ adds another layer of
understanding by highlighting the disparities in health, which could
explain some of the observed differences in the Ladder Scores across
different regions.
This step will establish a correlation between happiness and health,
and provides a more holistic view of regional well-being. The analysis
also helps reveal whether regions with higher Ladder Scores also have
longer healthy life expectancy, providing further evidence for the
relationship between health and happiness.
Analysis: The bar plot shows that Western Europe has the highest
average Healthy Life Expectancy, which aligns with its high Ladder
Scores observed in the box plots. This suggests that the high happiness
levels in Western Europe could be partly attributed to better health and
longer life expectancy. Likewise, North America and ANZ also have a high
average Healthy Life Expectancy, which corresponds to the relatively
high Ladder Scores in this region. This supports the notion that regions
with higher life expectancies tend to have higher happiness levels.
However, Sub-Saharan Africa has the lowest Healthy Life Expectancy,
which is consistent with its low Ladder Scores observed in the box
plots. The combination of poor health and low life expectancy could be
significant factors contributing to the lower Ladder Scores/ happiness
levels in this region.
In summary, the analysis of Healthy Life Expectancy provides crucial
context for understanding the regional disparities in ladder scores
(happiness) observed in the earlier steps. Regions with higher life
expectancy generally exhibit higher ladder scores or happiness levels,
reinforcing the importance of health as a key component of
well-being.
3.10. Analyze global happiness scores over entire time period
In the previous steps, the analysis focused on regional trends and
correlations between happiness, health, and other socio-economic
factors. These analyses were primarily region-based, offering a
high-level view of the data. The interactive choropleth map in this step
provides a dynamic perspective on how happiness has changed over time
across different countries, thereby offering insights into the overall
trends in global happiness. This map offers a different perspective by
emphasizing geographic distribution and temporal changes. By introducing
the dimension of time (in years), the choropleth map enables an analysis
of how happiness scores have evolved over time within specific
countries.
Pre-processing: When the analysis moves to a country-level
comparison, particularly when visualizing data on a map (such as a
choropleth map), the accuracy and consistency of country names become
crucial. So, this data cleaning step is performed to ensure that the
‘Country.name’ column in the dataset is free of any encoding issues,
non-ASCII characters, and other problematic strings that could cause
errors during analysis or visualization. By setting the locale,
converting text encoding, handling missing values, and removing invalid
characters, the dataset is prepared for accurate and error-free
plotting.
## [1] "en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8"
## [1] 0
Analysis: This interactive choropleth map provides the flexibility to
observe the trends in Ladder scores across time (2005 - 2023) over
different countries. In 2005, Denmark had the highest happiness score,
followed by other happy countries such as Canada, Australia, Saudi
Arabia, Spain, France, Sweden, the UK, and Venezuela. Conversely,
Romania had the lowest happiness score, with Hungary, Iran, Pakistan,
Poland, and Egypt also exhibiting low scores. In 2006, although most of
the countries observed in the previous year are not visible, Venezuela
still maintains a relatively high happiness score, albeit slightly lower
than the previous year. Togo and Benin in West Africa have some of the
lowest happiness scores, with 3.20 and 3.33, respectively. From 2007 to
2010, Canada consistently displayed high happiness scores, making it one
of the happiest countries. Denmark, despite experiencing fluctuations,
continued to hold the top spot until 2010 with a Ladder Score of
7.7.
However, by 2015, Denmark dropped to the third position on the list
of happiest countries, with Norway and Switzerland taking the first and
second positions, respectively. Liberia and Yemen emerged as the least
happy countries that year. In 2021, Finland topped the list with a
happiness score of 7.79, followed by Denmark and Iceland, while
Afghanistan remained at the bottom with a score of 2.43. In 2022,
Afghanistan’s Ladder Score further decreased to 1.28, keeping it at the
bottom of the list, while Finland maintained its position at the top
with a score of 7.72. Although Afghanistan’s Ladder Score improved
slightly to 1.46 in 2023 compared to 2022, it still ranked lowest on the
happiness scale. Finland, for the third consecutive year since 2021,
remained the happiest country.
3.11. Comparative Analysis of Happiness Scores - Identify Happiest
and Unhappiest Countries
After the choropleth map, which provided a temporal and geographical
overview of the Ladder scores across countries over time, the
comparative analysis of happiness scores was performed to pinpoint
specific countries that consistently rank as the happiest and unhappiest
based on their cumulative Ladder scores. It transitions from broad,
dynamic visualizations to a more focused analysis, allowing for the
identification of specific outliers (both positive and negative) in
terms of happiness.
Analysis of Bar Plots: The bar plot reveals the top three happiest
and unhappiest countries based on their cumulative Ladder scores.
Denmark, Sweden, and Canada have the highest cumulative happiness
scores, indicating that these countries have maintained consistently
high happiness levels over the years. While Guyana, Cuba, and Maldives
have the lowest cumulative happiness scores, suggesting that these
countries have consistently low happiness levels across the dataset’s
timespan. These findings differ from year-by-year analyses shown in the
choropleth map, where countries might temporarily rise or fall in
happiness rankings due to short-term factors that might not affect their
long-term cumulative scores.
In summary, these bar plots show a wide contrast in the happiness
scores between the top 3 happiest and unhappiest countries, and this
sets the stage for further analysis into the factors driving these
differences.
3.12. Check for Normality
After identifying the happiest and unhappiest countries, it is
essential to check the normality of the entire dataset and the subset of
happiest and unhappiest countries to make appropriate selection of the
statistical test to be performed. Understanding whether the data is
normally distributed will guide the choice between parametric and
non-parametric tests. And a Shapiro-Wilk normality test on the entire
dataset ensures that the subsequent analysis is accurate and reliable.
If the data is normally distributed, parametric tests such as the
T-test, which have higher statistical power, can be used. If the data is
not normally distributed, non-parametric tests would be more
appropriate.
##
## Shapiro-Wilk normality test
##
## data: ds_2023_clean$Life.Ladder
## W = 0.98964, p-value = 4.935e-12
##
## Shapiro-Wilk normality test
##
## data: top_countries$Total_Life_Ladder
## W = 0.70486, p-value = 0.006953
Analysis: Since p-value is way less than 0.05 in the entire dataset
and the combined dataset of the subsets - happiest and unhappiest
countries, these datasets do not follow a normal distribution. In case
of combined datasets, this could be because of merging two datasets that
have very different central tendencies (means). The happiest countries
have much higher Life Ladder scores compared to the unhappiest
countries, which might lead to a skewed distribution when combined. This
could in turn cause the Shapiro-Wilk test on the combined data to
indicate non-normality.
However, focusing on the normally distributed subsets - the happiest
and unhappiest countries individually, enables the advantages of
parametric testing, such as greater power and precision, to be
leveraged. This approach is crucial for understanding the sharp
contrasts between the happiest and unhappiest countries. By isolating
and analyzing these subsets, more meaningful comparisons between the
extremes of the happiness spectrum can be made, providing clearer and
more actionable insights.
##
## Shapiro-Wilk normality test
##
## data: top_3_happiest$Total_Life_Ladder
## W = 0.88385, p-value = 0.3358
##
## Shapiro-Wilk normality test
##
## data: top_3_unhappiest$Total_Life_Ladder
## W = 0.93768, p-value = 0.5181
Analysis of Shapiro-Wilk Normality test: The p-value is greater than
0.05 for Happiest and Unhappiest countries (p-value = 0.3358 and p-value
= 0.5181), indicating that the data for both the happiest and unhappiest
countries individually are normally distributed.
Thus, the data of Happiest and Unhappiest countries can be tested
individually using a parametric test (T-test) since these two data when
considered individually are normally distributed. T-test can be used to
compare the mean happiness scores (Total Life Ladder) between the
happiest and unhappiest countries.
3.13. Perform T- Test
In order to statistically validate the differences between the
happiest and unhappiest countries, a T-test is performed.
##
## One Sample t-test
##
## data: top_3_happiest$Total_Life_Ladder
## t = 65.99, df = 2, p-value = 0.0002296
## alternative hypothesis: true mean is not equal to 0
## 95 percent confidence interval:
## 125.2224 142.6909
## sample estimates:
## mean of x
## 133.9567
##
## One Sample t-test
##
## data: top_3_unhappiest$Total_Life_Ladder
## t = 23.36, df = 2, p-value = 0.001828
## alternative hypothesis: true mean is not equal to 0
## 95 percent confidence interval:
## 4.516601 6.556066
## sample estimates:
## mean of x
## 5.536333
Analysis of the test: The Happiest Countries have a high mean
happiness score (Total Life Ladder) of 133.96, which is statistically
significant, as indicated by the low p-value (0.0002296) and a high
t-value (65.99). This suggests that these countries consistently score
very high on the happiness scale. However, the Unhappiest Countries have
a much lower mean happiness score of 5.54, also statistically
significant (p-value: 0.001828, t-value: 23.36). This confirms that
these countries are at the opposite end of the happiness spectrum, but
still have a score that is statistically above zero, indicating some
level of happiness.
In summary, the T-test results show that the happiest countries have
a significantly higher mean happiness score (Total Life Ladder) compared
to the unhappiest countries. The T-test confirms that the differences
between these two groups are not just due to random chance but are
statistically significant.
3.14. Correlation Analysis of Factors Influencing Happiness
After confirming the significant difference in happiness scores
between the happiest and unhappiest countries, the next step is to
understand what drives these differences. This step analyzes the
correlation between various factors (such as GDP per capita, social
support, etc.) and the Life Ladder score for the happiest and unhappiest
countries. Understanding these relationships will help uncover the key
factors influencing happiness in these countries.
## [1] "Canada" "Cuba" "Denmark" "Guyana" "Maldives" "Sweden"
## # A tibble: 3 × 3
## Country.name Factor Correlation
## <chr> <chr> <dbl>
## 1 Canada Healthy.life.expectancy.at.birth -0.836
## 2 Denmark Healthy.life.expectancy.at.birth -0.757
## 3 Sweden Social.support 0.354
Analysis: The analysis reveals that different factors significantly
impact happiness across these countries. In Canada and Denmark, the
unexpected negative correlation between Healthy Life Expectancy and
happiness suggests that longevity alone does not guarantee happiness;
other factors related to the quality of life, healthcare systems, and
social conditions might play a crucial role. In contrast, Sweden’s
positive correlation between Social Support and happiness underscores
the importance of social connections and community support in enhancing
well-being. These findings highlight the complex and multifaceted nature
of happiness, demonstrating that it is influenced by a variety of
factors that can differ significantly from one country to another.
Understanding these relationships helps in identifying key areas for
policy intervention and improvement to enhance overall well-being in
different regions.
3.15. Regional Correlation Analysis
After examining the correlation factors in the happiest and
unhappiest countries, it’s important to expand the analysis to a
regional level. This step explores how different factors correlate with
happiness scores across various regions, helping to identify
region-specific trends and influences.
Analysis: The heatmap reveals that the relationship between happiness
and various socio-economic factors varies significantly across regions.
Western Europe, North America, and ANZ consistently show positive
correlations across multiple factors, suggesting that in these regions,
improvements in economic and social conditions lead to higher happiness.
On the other hand, regions like East Asia and Sub-Saharan Africa show
weaker or even negative correlations with some factors, indicating that
the drivers of happiness may be more complex or less directly linked to
these specific variables. This heatmap underscores the importance of
regional context in understanding the factors that influence happiness
and suggests that a one-size-fits-all approach may not be appropriate
when assessing global happiness metrics. It highlights the need for
region-specific strategies to improve happiness and well-being.
3.16. Analysis of happiness score vs socio-economic factors
The heatmap in the previous step underscores the importance of
regional context in understanding the factors that influence happiness
and suggests that a one-size-fits-all approach may not be appropriate
when assessing global happiness metrics. It highlights the need for
region-specific strategies to improve happiness and well-being. To
further explore these regional trends and the relationships between
happiness and various socio-economic factors, scatter plots with
regression lines are created. This step provides a more detailed
visualization, allowing for the examination of linearity and the
strength of these relationships, thereby confirming and complementing
the findings from the correlation analysis.
##
## Attaching package: 'gridExtra'
## The following object is masked from 'package:dplyr':
##
## combine
## Warning: Removed 3 rows containing non-finite outside the scale range
## (`stat_smooth()`).
## Warning: Removed 3 rows containing missing values or values outside the scale range
## (`geom_point()`).
## Warning: Removed 3 rows containing non-finite outside the scale range
## (`stat_smooth()`).
## Warning: Removed 3 rows containing missing values or values outside the scale range
## (`geom_point()`).
## Warning: Removed 3 rows containing non-finite outside the scale range
## (`stat_smooth()`).
## Warning: Removed 3 rows containing missing values or values outside the scale range
## (`geom_point()`).
## Warning: Removed 3 rows containing non-finite outside the scale range
## (`stat_smooth()`).
## Warning: Removed 3 rows containing missing values or values outside the scale range
## (`geom_point()`).
## Warning: Removed 3 rows containing non-finite outside the scale range
## (`stat_smooth()`).
## Warning: Removed 3 rows containing missing values or values outside the scale range
## (`geom_point()`).
## Warning: Removed 3 rows containing non-finite outside the scale range
## (`stat_smooth()`).
## Warning: Removed 3 rows containing missing values or values outside the scale range
## (`geom_point()`).
Analysis: The scatter plots with regression lines reveal the
following relationships between happiness (Ladder score) and
socio-economic factors:
GDP per Capita: A strong positive relationship, indicating that
higher economic output per person is linked to greater happiness. Social
Support: A positive correlation, suggesting that stronger social
networks contribute significantly to higher happiness levels. Life
Expectancy: A positive relationship, emphasizing that longer, healthier
lives are associated with increased happiness. Freedom to Make Choices:
A positive correlation, showing that greater autonomy and freedom in
life choices lead to higher happiness. Generosity: A weaker correlation,
indicating that while generosity contributes to happiness, it is less
impactful than other factors like economic stability or social support.
Perception of Corruption: A more complex and weaker positive
relationship, suggesting that the impact of corruption on happiness
varies and may be influenced by other mitigating factors.
3.17. Interactive dashboard to investigate global impact of
socio-economic factors on happiness
To enhance the depth of analysis, an advanced interactive
visualization is created. This interactive plot allows for a more
granular exploration of how different factors influence happiness across
regions. By enabling the examination of region-specific correlations,
this visualization provides an opportunity to interact with the data
dynamically, offering deeper insights into the complex relationships
between happiness and its drivers.
Note: In slidey mode, every click, including those on dropdown menus,
may advance to the next slide. After selecting variables from the
dropdown, please manually navigate back to the page to explore the
interactive dashboard.
Analysis: This interactive visualization provides a dynamic tool for
exploring how different factors correlate with happiness across various
regions. By allowing users to filter and compare regions, it offers a
more detailed understanding of how regional contexts influence the
relationship between different socio-economic factors and happiness.
Section 4: Conclusion and
Future Scope
This exploratory data analysis (EDA) has provided a comprehensive
examination of the factors influencing global happiness. Through a
combination of statistical tests and visualizations, key insights into
the drivers of happiness have been uncovered, highlighting the
importance of regional context and the significant role of economic and
social factors. The analysis also reveals the complexity of the
relationships that contribute to overall well-being.
The advanced interactive visualization developed in this project
serves as a powerful tool for further exploration, offering a dynamic
way to delve deeper into the data. However, while the current analysis
provides valuable insights, it also opens avenues for further
exploration and refinement. Incorporating more sophisticated statistical
techniques, such as multivariate regression or machine learning, could
yield even deeper insights into the complex relationships between
socio-economic factors and happiness.
Additionally, the interactive dashboard could be enhanced with
features like multi-factor analysis, real-time updates, and predictive
modeling to offer more dynamic and actionable insights.
This thorough EDA not only lays the foundation for more targeted
research but also supports interventions aimed at improving happiness on
both a global and regional scale.